Mining Weighted Preferred Traversal Patterns in Fuzzy Environment

摘要

In this paper, we proposed an effective algorithm to mine preferred traversal patterns from web logs. The importance of a web page, browsing times and the time duration on a web page are considered as three important factors to reveal users' interest. Linguistic term reflecting importance of a web page is characterized as a fuzzy linguistic variable. It was transformed as corresponding fuzzy weight expressing importance degree of the web page. Besides, we characterize the time duration on a web page as a fuzzy variable. By the comparison of their expected values, it is clear that users show different interest in these web pages. Another factor, browsing times, is considered as relative frequency. Based on the presented consideration, a new concept of weighted fuzzy preference is proposed to replace the traditional confidence during the mining process. And an novel algorithm based on Frequent Link and Access Tree (FLAAT) is developed to mine preferred traversal patterns. In the end an example is provided to clearly illustrate the proposed approach.